Quality assurance of Cyberknife robotic stereotactic radiosurgery using an angularly independent silicon detector

Abstract Purpose The aim of this work was to evaluate the use of an angularly independent silicon detector (edgeless diodes) developed for dosimetry in megavoltage radiotherapy for Cyberknife in a phantom and for patient quality assurance (QA). Method The characterization of the edgeless diodes has been performed on Cyberknife with fixed and IRIS collimators. The edgeless diode probes were tested in terms of basic QA parameters such as measurements of tissue‐phantom ratio (TPR), output factor and off‐axis ratio. The measurements were performed in both water and water‐equivalent phantoms. In addition, three patient‐specific plans have been delivered to a lung phantom with and without motion and dose measurements have been performed to verify the ability of the diodes to work as patient‐specific QA devices. The data obtained by the edgeless diodes have been compared to PTW 60016, SN edge, PinPoint ionization chamber, Gafchromic EBT3 film, and treatment planning system (TPS). Results The TPR measurement performed by the edgeless diodes show agreement within 2.2% with data obtained with PTW 60016 diode for all the field sizes. Output factor agrees within 2.6% with that measured by SN EDGE diodes corrected for their field size dependence. The beam profiles’ measurements of edgeless diodes match SN EDGE diodes with a measured full width half maximum (FWHM) within 2.3% and penumbra widths within 0.148 mm. Patient‐specific QA measurements demonstrate an agreement within 4.72% in comparison with TPS. Conclusion The edgeless diodes have been proved to be an excellent candidate for machine and patient QA for Cyberknife reproducing commercial dosimetry device measurements without need of angular dependence corrections. However, further investigation is required to evaluate the effect of their dose rate dependence on complex brain cancer dose verification.

that employs multiple narrow beams to deliver conformed and precise high radiation dose to the target from different directions in single or few fractions. 1,2 It requires an accurate target localization and identification which can be achieved by physical stereotactic immobilization devices registering patient to a fixed frame (e.g., Gammaknife) or by imaging-guided methods (such as Cyberknife Synchrony). 3 Due to the small beam size and precise conformation of dose distribution, SRS treatment can reduce radiotoxicity to normal tissues and organs at risk and improve the probability of local tumor control. 4 It is used often for intracranial (brain tumor) and recently extracranial lesions such as spine and breast tumors.
The small treatment volume sizes that are used in SRS introduce several dosimetric challenges for quality assurance (QA) which are not observed in standard conformal radiotherapy. Most predominant challenges are related to the dimensions of the detectors relative to the radiation field size which leads to a volume averaging effect and the fluence perturbation caused by the materials adopted for fabrication of the devices. Perturbation is created due to the variety of stopping power ratios of the materials composing the sensitive volume and surrounding packaging of the detector relative to water and consequently the alteration of the detector response. [5][6][7][8][9][10][11][12] Due to these effects, the uncertainty in small field dosimetry is significantly higher and errors are notably larger than in dosimetry of traditional radiotherapy field sizes. In nonisocentric radiation delivery modalities, all these effects must be combined with the angular dependence of the dosimetry devices which cannot be easily mitigated using a correction factor based on the relative position of the linac gantry. Ideally, the detectors used for QA in robotic SRS equipment such as Cyberknife should be energy, dose rate, and angular independent. In addition, they should have the ability to obtain high spatial resolution measurement without perturbing the radiation beam. 4,[12][13][14] Although ionization chambers are considered a reference standard in radiotherapy dosimetry, 4,15.16 the relative large size of the sensitive volume introduces severe volume averaging effects for the smallest field sizes which overestimate the penumbra of the field and underestimate the output factor. 4,17 Additionally, mini chambers suffer from reduction in their sensitivity and increased noise level due to their small sensitive volume size. 4 Radiochromic films have been widely used in small field dosimetry because of their near water-equivalent material and the suitability for measuring dose profiles with high spatial resolution. 18,19 They are also angularly independent but suffer from lack of reproducibility which depends on processing conditions and procedure. Diamond detectors have been of high interest in small field measurement recently for their near tissue equivalence in a photon beam, high spatial resolution, and realtime readout. 1 However, they are expensive and exhibit dose rate dependence 5,12 and interdevice reproducibility. Silicon diodes are one of the most common detectors adopted for small field dosimetry. The relatively low average ionization energy required to produce an electron-hole pair (3.6 eV) and its density make silicon diodes very sensitive and very small sensitive volumes can be manufactured. 20 The mass collision stopping power ratio of electrons for silicon-water makes silicon diodes almost completely energy independent for MV range energies. 20 However, the application of silicon diodes in a small field measurement, especially in nonisocentric noncoplanar and flattering filter free (FFF) modalities like Cyberknife, is limited by directional and dose rate dependence.
The angular dependence of silicon diodes results from their geometry and construction; directionality depends also on the energy of incident beam, field size, and the back scattering from the packaging material creating variations in sensitivity up to 25% with angle of incidence. 21 There have been several reported solutions to overcome detectors responses anisotropy. One solution has been introduced by Jursinc et al. by adding a thin copper disk to the top side of the diode used in the MapCHECK device which has decreased the angular dependence from ±10% to ±1.25%. However, this solution increased the perturbation of radiation beam due to the addition of the copper material which makes the correction factors depend on the beam energy. 21,22 Westermark et al. proposed another solution by coupling two diodes back-to-back similar to the approach used in MOSFETs. 12,23 The combination of two diodes is found to mitigate the angular dependence to just ±3%, but the double mass of the diodes makes this solution unsuitable for small field dosimetry due to a large beam perturbation. 24 Several correction factors based on the solutions of directional dependence have been adopted by many research groups and companies for the optimization of commercially available silicon diodes used in QA devices such as the Delta4 (ScandiDos, Uppsala, Sweden), ArcCHECK (SunNuclear, Melbourne, FL, USA), and ion chambers arrays such as I'mRT MatriXX (IBA Dosimetry, Schwarzenbruck, Germany). This solution requires the measurement of the angle of the beam with respect to the detector and applying a correction factor for each angle. This approach is not implemented yet for robotic radiotherapy delivery ALHUJAILI ET AL.
| 77 modalities such as Cyberknife SRS which requires a characterization in almost the whole solid angle.
The Centre for Medical Radiation Physics (CMRP) has proposed a solution to overcome the issue of the angular dependence of silicon diodes by replacing the conventional semiconductor planar structure with a design of the junction close to being a symmetrical three-dimensional (3D) shape and adopting an innovative diode packaging approach. The technology proposed is called "edgeless" or "active edge" detector. This fabrication technology has been developed by the VTT Technical Research Centre of Finland Micro and Nanoelectronics (Finland) within the framework of the international collaboration MEDIPIX, and its application in radiotherapy dosimetry, in combination with the "drop in" packaging technology, is proposed by CMRP. 21 The basic characterization of the edgeless detectors for dosimetry in external beam radiotherapy is described in Petasecca et al. 21 The aim of this study was to evaluate the application of the angularly independent "edgeless" detectors as a QA tool for robotic SRS modalities such as Cyberknife ® by testing the diodes for routine dosimetric QA and by delivering three full patient plans to a lung phantom which is stationary or moving with a breathing pattern recorded from four-dimensional CT for the same patients. In this work, absolute and relative measurements including a field size factor, dose off-axis profiles, and tissue-phantom ratio (TPR) have been performed. Measurements were also performed for comparison using PTW 60016, SNC Edge, PinPoint ionization chamber, and Radiochromic EBT3 Films.

2.A | Edgeless detectors
The edgeless detectors are fabricated using a lateral implantation technique instead of a standard planar semiconductor fabrication processes. The lateral implantation produces a 3D p-n junction (or ohmic contact) surrounding the die that is leading to full charge collection. Although the edgeless technology allows for processing of both p-and n-type substrates, in this work, the devices adopted are only n-type, with the top side junction being p + −n and the lateral junction n + −n. The diodes have dimensions of 0.5 × 0.5 × 0.5 mm 3 [ Fig. 1(a)] and are packaged using the "dropin" proprietary CMRP technology [ Fig. 1(b)]. The packaging is water tight and allows for measurements in a water phantom. The edgeless diodes are readout by a custom-designed acquisition system based on a commercially available multichannel electrometer named TERA (Tera Foundation, Turin, Italy) which is described in detail by Mazza et al. 25

2.B | CyberKnife ® robotic stereotactic radiosurgery systems
CyberKnife is a SRS machine that consists of a portable linear accelerator mounted on an industrial robotic arm (manipulator). By utilizing a set of collimators and a sophisticated imaging-based tracking system, CyberKnife can produce small, noncoplanar radiation beams and deliver them to a target located near to critical structures. There are two different collimation systems: one system is a collection of fixed collimators (cones) which are manufactured from metallic material with 12 different diameters (from a diameter of 5 to 60 mm).
The second system is the Iris TM collimator, a variable aperture dia- dosimetric measurements with the edgeless diodes have been performed on the G4 machine, the phantom study measurements were performed using both the M6 and G4 generations Cyberknife.

2.C.1 | Timber phantom
Cyberknife is also used for clinically suitable lung lesions, particularly when the lesion is in proximity to organs at risk thanks to its capability to track the motion of the target accurately. 27 In order to test the edgeless detectors for patient-specific QA, two timber phantoms have been manufactured to mimic a lung with and without an internal lesion. The heterogeneous phantom which presents the internal lesion is composed of two cubic blocks of timber (with a density of approximately 0.3 g/cm 3 ) with one hemisphere of solid water in each block positioned at the center of the phantom. The solid water insert mimics a lesion of a diameter of approximately 2 cm inside the lung.
The detectors are positioned in between the timber blocks with one hemisphere above and below, to form a spherical lesion with 1 mm gap (Fig. 2). The heterogeneous phantom has been manufactured at the University of Wollongong mechanical workshop and has dimensions of 9.45 × 10 × 14.7 cm 3 with two slabs of solid water, 2 cm thick above and below the timber blocks to mimic the attenuation from the chest wall muscles and backscattering from the back muscles. In this work, we used also a homogenous version of the timber phantom with the same dimensions and configuration of the heterogeneous phantom but without the internal lesion.

2.D | Verification of response angular dependence for noncoplanar irradiations
A key characteristic of the edgeless detectors is the angular independence, particularly important in Cyberknife due to its intrinsic noncoplanar radiation delivery. The edgeless detectors have been characterized in terms of angular dependence in cross-and in-plane delivery in Ref. [21] with variation in the response within ±2% for angles between ±90 degree. In this work, we performed also a delivery of the radiation in a plane at 45 degrees between the cross-and  in-plane directions. Irradiation has been performed by a Varian Truebeam with 6 MV flattening filter free (FFF) with the couch set at 45 degree and a cylindrical Perspex phantom with the sample placed at isocenter at 15 cm depth. The field size adopted for this test is 10 × 10 cm 2 collimated using the jaw collimators. We did not perform the angular dependence on Cyberknife because the free-positioning system of the machine does not allow a fine control of the angle between the beam and the plane of the couch, while the Truebeam alignment system allows for a more accurate positioning of the phantom and control of the gantry around the isocenter.

2.E | Linearity and calibration factor
Calibration and verification of the response linearity of the edgeless diodes were performed under reference calibration conditions with the Cyberknife head perpendicular to the phantom at source to detector distance (SDD) of 800 mm and using the fixed cone of 60 mm diameter as suggested by the IAEA-493. 28 The detectors have been placed at a depth of 1.5 cm and calibrated by irradiating each device in increments of 100 cGy up to a total accumulated dose of 400 cGy. Each irradiation step has been repeated three times to evaluate the repeatability of the measurement.

2.G | Field size factor measurement
Field size factor is a parameter which must be characterized for each machine and collimation system adopted. The measurement of the field size factor was carried out in a medium size MP3 motorized water tank at the Cyberknife G4. The edgeless diode was attached to a plastic holder allowing it to be remotely controlled for 3D movement in the water phantom with a step resolution of 0.1 mm.
The diode was placed at a depth 15 mm and its lateral position was adjusted remotely to obtain maximum signal corresponding to the center of the radiation field from the collimator. The alignment procedure was repeated for each filed size.

2.I. | Beam profile measurement
Profile measurements were performed with the diode embedded in a solid water phantom equipped with a two-axis stepper motor stage. After the alignment, performed with the same procedure adopted for OF and TPR measurements, the Cyberknife head was kept static with the radiation beam perpendicular to the phantom surface. The diode was moved across the beam at constant speed (a margin of a few centimeters ensured speed stabilization). The radiation field sizes measured were 5, 10, 30, and 60 mm collimated by Iris collimators at an SDD of 800 mm and a depth in solid water of 15 mm.

2.J | Patient-specific QA measurement
In order to assess the performance of the edgeless detectors in semi-analytic method using experimental data such as off-axis ratio, TPR, and output factor to calculate the dose kernel and the effective path length to correct for heterogeneities. 29 The second method is Monte Carlo which adopts a virtual source (phase space file of the linac head) to calculate the dose. 29 Three plans of uniform coverage were created using the RTrac method.
Plan 1 and Plan 2 were created using the heterogeneous phantom (Fig. 3)    Each set (called a node) contains one or more beams which are delivered to the target through unique linac head positions in space. The full set of nodes is called path set which is usually constructed and optimized by the TPS (with no or marginal control from the operator) to deliver the plan. The details of the plans are summarized in Table 2.

2.K | Patient-specific QA measurement using EBT3
Gafchromic EBT3 film was used as benchmark for the patient-specific QA measurements. The film was cut into 7 × 7 cm 2   3.C | Dose per pulse dependence 3.D | Field size factor Figure 6 show the field size factors measured by the edgeless detector with IRIS collimator. The x-axis shows the diameter of the equivalent circular field size ranging from 5 to 60 mm at SDD of 650, 800, and 1000 mm. The response of the edgeless diodes has been compared to SNC EDGE diode. The overresponse of the SNC EDGE diodes in the smallest fields has been corrected for by applying the corresponding field correction factors reported by Francescon. [30][31][32] The edgeless diodes show an agreement with SNC EDGE diodes in the field size range of 25 to 60 mm with discrepancies within ±1%, while at smaller field sizes from 5 to 20 mm, discrepancies do not exceed ±2.6%.   Table 3 shows full width half maximum (FWHM) and penumbra width (80%-20%) of the profiles which have been obtained by using an interpolation-shape-preserving fit (with a resolution step of 0.01 mm). Figure 8 and Table 3 show an agreement between the FWHM recorded by the edgeless and the SNC EDGE diodes within 2.3% for all the beam profiles and the discrepancies in penumbra width are within 0.148 mm.

3.F | Beam profiles measurements
3.G | Patient-specific QA measurement Patient plans were also simulated and delivered to a lung phantom with four edgeless diodes placed across the gross target volume.
The differences between patient-specific QA measurements with the edgeless diodes were within 4.72% when compared to TPS, for all the phantom configurations. These preliminary results are limited in terms of type of plan delivered and clinical scenarios adopted but demonstrate that the edgeless diodes are a valuable technology also for patient QA, providing a real-time dosimetry evaluation also for noncoplanar radiotherapy modalities, without requiring a correction factor for angular dependence, even when organized in an array of multiple single diodes.

Sultan Alhujaili is supported by Aljouf University, College of Applied
Medical Sciences, Radiology and Medical Imaging Department.

CONF LICTS OF INTEREST
The authors have no relevant conflicts of interest to disclose.